EstimatorQNN mismatch between installed ML 0.8.0 and its GitHub version

[ironfrown]

Environment

• 0.8.0:
• 3.1.10:
• Ubuntu 22.04.5 LTS:

What is happening?

ML 0.8.0 EstimatorQNN with V2 primitives does not accept pass_manager option, but the sources on GitHub allow this (instead an old num_virtual_qubits is being accepted).

How can we reproduce the issue?

Install ML 0.8.0, create a simple circuit, call EstimatorQNN with V2 estimator and option: pass_manager and/or num_virtual_qubits

What should happen?

Installed and GitHub versions should be consistent

Any suggestions?

No response

[edoaltamura]

Hi @ironfrown, the mismatch appears to be related to #866.

Both these issues are scheduled to be fixed in v.0.8.1, expected for this week. I propose keeping this issue open until the release.

For reference, the new EstimatorQNN implementation was updated in #855 and #870.

[edoaltamura]

Does this issue persist in version 0.8.1 @ironfrown?

[edoaltamura]

The mismatch was resolved in version 0.8.1, and

ML 0.8.0 EstimatorQNN with V2 primitives does not accept pass_manager option, but the sources on GitHub allow this (instead an old num_virtual_qubits is being accepted).

was fixed in #870.